Detection of Multidrug-Resistant Salmonella enterica Serotype typhimurium DT104 Based on a Gene Which Confers Cross-Resistance to Florfenicol and Chloramphenicol

Abstract

Salmonella enterica serotype typhimurium (S. typhimurium) DT104 (DT104) first emerged as a major pathogen in Europe and is characterized by its pentadrug-resistant pattern. It has also been associated with outbreaks in the United States. The organism typically carries resistance to ampicillin, chloramphenicol, streptomycin, sulfonamides, and tetracycline. The mechanism of chloramphenicol resistance in DT104 was determined by producing antibiotic-resistant Escherichia coli host strain clones from DT104 DNA. DNA from chloramphenicol-resistant clones was sequenced, and probes specific for the genes flo_{S. typhimurium} (flo_St), int, invA, and spvC were produced for colony blot hybridizations. One hundred nine Salmonella isolates, including 44 multidrug-resistant DT104 isolates, were tested to evaluate the specificities of the probes. The gene flo_St, reported in this study, confers chloramphenicol and florfenicol resistance on S. typhimurium DT104. Florfenicol resistance is unique to S. typhimurium DT104 and multidrug-resistant S. typhimurium isolates with the same drug resistance profile among all isolates evaluated. Of 44 DT104 isolates tested, 98% were detected based on phenotypic florfenicol resistance and 100% had the flo_St-positive genotype. Resistances to florfenicol and chloramphenicol are conferred by the gene flo_St, described in this paper. Presumptive identification of S. typhimurium DT104 can be made rapidly based on the presence of the flo_St gene or its resulting phenotype.

Salmonella enterica serotype typhimurium (S. typhimurium) definitive type 104 (S. typhimurium DT104 or DT104) is an increasingly common multiple-antibiotic-resistant strain of Salmonella that has rapidly emerged as a world health problem (5, 21). The DT is based on phage typing of the organism. DT104 is characterized by chromosomal resistance to ampicillin (A), chloramphenicol (C), streptomycin (S), sulfonamides (Su), and tetracycline (T) and is commonly referred to as having resistance (R) type ACSSuT (20).

Antibiotic resistance is increasing among many bacterial species and is rapidly becoming a major world health problem (5, 16). The Centers for Disease Control and Prevention (Atlanta, Ga.) maintain several programs designed to monitor antimicrobial resistance. As part of the 1995 National Salmonella Antimicrobial Resistance Study, the Centers for Disease Control and Prevention serotyped 3,903 Salmonella isolates and determined that 976 (25%) were S. typhimurium. Approximately 28% (275 of 976) of the S. typhimurium isolates had the R type ACSSuT compared to just 7% in 1990 (7). Two other antibiotic resistance monitoring programs, the National Antimicrobial Resistance Monitoring System and Periodic Surveys of Antimicrobial Drug Resistance in Sentinel Counties, have reported similar sharp increases in multidrug-resistant Salmonella (5).

The mechanisms for resistance to sulfonamides, streptomycin, and ampicillin in DT104 have been described previously (15, 18). DT104 contains at least two integrons, one containing the aminoglycoside resistance gene cassette ant(3")-Ia, which encodes resistance to streptomycin, and one containing a β-lactamase gene cassette, pse-1, which encodes resistance to ampicillin (15, 18). A gene coding for sulfonamide resistance (sul-1) was found in the 3′ conserved sequences of both integrons. The mechanisms for resistance to tetracycline and chloramphenicol have not been reported.

This study reports the discovery of a new gene called flo_{S. typhimurium} (flo_St) and the use of flo_St-based gene probe to detect S. typhimurium DT104. Our objectives were to ascertain the mechanism of chloramphenicol resistance in DT104, to determine if DT104 is florfenicol resistant, and to discern if chloramphenicol and florfenicol resistances are linked. In addition, we assessed the utility of the florfenicol-resistant phenotype and the flo_St genotype as diagnostic tools for identification of S. typhimurium DT104. Finally, we examined DT104 isolates to determine the frequency with which the integrons invA and spvC occurred, and we sought to determine if a diagnostically useful relationship between these genotypes and DT104 exists.

MATERIALS AND METHODS

Lambda library construction and screening of clones.

Standard recombinant DNA procedures were performed as described by Sambrook et al. (17). Chromosomal DNA from S. typhimurium DT104 cattle isolate 152N17 was prepared for construction of a lambda library by partial SauIIIA (New England Biolabs, Beverly, Mass.) digestion. The partial digestion yielded DNA fragments ranging from 500 bp to 12 kbp. Digested DNA was extracted with phenol-chloroform and ethanol precipitated.

The SauIIIA-digested DNA was ligated into Lambda Zap Express Vector (Stratagene, La Jolla, Calif.) digested with BamHI according to instructions provided by the manufacturer. The library was then packaged into lambda phage (MaxPlax Packaging Extract; Epicentre Technologies, Madison, Wis.) and used to transduce the Escherichia coli XL1-Blue host strain. The vector was excised from E. coli XL1-Blue in the form of plasmid pBK-CMV and used to transform the E. coli XLOLR host strain according to the instructions provided in the Lambda Zap Express Vector kit. Plasmid preparations from colonies that grew on Luria-Bertani agar (Difco Laboratories, Detroit, Mich.) supplemented with chloramphenicol (25 μg/ml) were made. Purified plasmid DNA was used to transform competent Library Efficiency E. coli DH5αMCR (Gibco-BRL, Grand Island, N.Y.) according to the manufacturer’s instructions. The transformed cells were screened for chloramphenicol resistance at a concentration of 25 μg/ml as described above, and plasmid was purified from chloramphenicol-resistant isolates. Plasmid pLB510 was determined to be chloramphenicol resistant and stable. Insert DNA from plasmid pLB510 was excised by digestion with EcoRI and SalI, yielding a 3.3-kb fragment. DNA was also excised by digestion with EcoRI and PstI, yielding a 2.1-kb fragment and a 1.2-kb fragment. Plasmid pLB510 was purified and insert DNA was sequenced by the dideoxy chain termination method at the University of Georgia Molecular Genetics Instrumentation Facility (Athens, Ga.) (MGIF).

Isolates for probe specificity and antibiotic resistance testing.

One hundred nine Salmonella isolates were selected from the banked culture collection which is part of the National Antimicrobial Susceptibility Monitoring System (22). Isolates were obtained from cattle, swine, chickens, turkeys, carcass rinses and washes, swine and cattle feeds, exotics (lizards, snakes, and iguanas), dogs, and cats from both clinically ill and nonclinical animals. All isolates were serotyped and tested for resistance to ampicillin, chloramphenicol, streptomycin, sulfonamides, tetracycline, and florfenicol. Testing, for all but florfenicol, was done with a semiautomated system (Sensititre; Accumed, Westlake, Ohio) according to the manufacturer’s directions. A custom-designed microtiter plate panel with the minimal inhibitory concentration format was used for all isolates. All isolates were maintained at −70°C. Florfenicol resistance was determined by disk diffusion assay on Mueller-Hinton agar plates (Schering-Plough Animal Health, Kenilworth, N.J.) according to the manufacturer’s instructions. Florfenicol and chloramphenicol MIC tests were conducted for selected Salmonella isolates by the microtiter broth dilution technique (8). S. typhimurium strains with R types characteristic of DT104 were phage typed. Serotyping and phage typing were conducted at the National Veterinary Services Laboratory, Ames, Iowa.

PCR probes and specificity for DT104.

The Salmonella isolates described above were tested with probes designed to detect the 2.1-, 1.2-, and 3.3-kb fragments from pLB510, the chloramphenicol and florfenicol resistance gene (flo_St), the integrase gene (int), the Salmonella invasion gene (invA), and the Salmonella virulence plasmid gene (spvC). A single colony of each Salmonella isolate was prepared by detergent lysis on nylon filters according to standard methods (17). Hybridization probes were purified as restriction endonuclease DNA fragments from recombinant plasmid pLB510 and/or synthesized as PCR amplification products (Table 1). The PCR-derived integrase probe was produced from E. coli SK1592(pDU202) with Int 1 and Int 2 primers specific for the integrase gene (10). The PCR-derived flo_St probe was produced from S. typhimurium DT104 isolate 152N17. Primers were designed (with GeneRunner version 3.04 [Hastings Software Inc.]) from nucleotide sequence of the gene flo_St (Table 1). Primers were synthesized by the MGIF with an ABI model 394 DNA synthesizer. The flo_St PCR product was sequenced by dideoxy chain termination to ensure its identity (MGIF). PCR-derived probes for invA and spvC were produced from S. typhimurium SR11 (19). All probes were labeled with digoxigenin deoxynucleoside triphosphates by either PCR or random primer extension (DIG High Prime labeling and detection starter kit; Boehringer Mannheim, Indianapolis, Ind.). Hybridizations and washes were carried out at high stringency (0.1% sodium dodecyl sulfate and 0.1% SSC at 68°C) (17). Positive and negative controls were included for all hybridization colony blots. The positive controls were pLB510 and S. typhimurium DT104 152N17 for the flo_St blot, E. coli SK1592(pDU202) for the int blot, and S. typhimurium SR11 for the invA and spvC blots. The negative controls for all blots were E. coli XL1-Blue, E. coli XLOLR, and E. coli DH5α; S. typhimurium SR11 and S. typhimurium pACYC184 also served as negative controls for the flo_St and int blots.

TABLE 1.

PCR primers

Name	Target	Sequence	Expected product size (bp)	Refer-ence
Flo 1	floSt	AATCACGGGCCACGCTGTATC	215	This study
Flo 2		CGCCGTCATTCTTCACCTTC
Int 1	int	CCTCCCGCACGATGATC	280	10
Int 2		TCCACGCATCGTCAGGC
Inv 1	invA	TTGTTACGGCTATTTTGACCA	521	19
Inv 2		CTGACTGCTACCTTGCTGATG
Spv 1	spvC	CGGAAATACCATCTACAAATA	669	19
Spv 2		CCCAAACCCATACTTACTCTG

Nucleotide sequence accession number.

The GenBank accession number of the flo_St gene is AF097407.

RESULTS

The following results pertain to the gene flo_St, which confers florfenicol and chloramphenicol resistance on S. typhimurium DT104. Nearly all R type ACSSuT S. typhimurium DT104 isolates also have resistance to florfenicol (Table 2; also see Table 4). Of 44 multidrug-resistant DT104 isolates tested, 43 were resistant to florfenicol and 1 had intermediate resistance (Table 2). The MIC of florfenicol ranged from 37.5 to 150 μg/ml, while chloramphenicol MICs ranged from 100 to greater than 200 μg/ml for the DT104 isolates tested.

TABLE 2.

Summary analysis of antibiotic resistance

Phenotypea	Total no. (%) positive
	Multidrug-resistant DT104b	Antibiotic-sensitive DT104c	Multidrug-resistant S. typhimurium not DT104d	ChlrSalmonella not typhimurium	ChlsS. typhimurium
Chlr	44 (100)	0 (0)	18 (100)	11 (65)	0 (0)
Chli	0 (0)	0 (0)	0 (0)	6 (35)	0 (0)
Chls	0 (0)	5 (100)	0 (0)	0 (0)	25 (100)
Flos	43 (98)	0 (0)	15 (83)	3 (18)	0 (0)
Floi	1 (2)	0 (0)	0 (0)	5 (29)	0 (0)
Flos	0 (0)	5 (100)	3 (17)	9 (53)	25 (100)
Total	44	5	18	17	25

^aChl, chloramphenicol; Flo, florfenicol. A superscript i indicates intermediate resistance. 

^bS. typhimurium DT104 isolates with R type ACSSuT, except for seven isolates with R type ACSuT. 

^cS. typhimurium DT104 isolates with no known antibiotic resistance. 

^dS. typhimurium isolates with R type ACSSuT or ACSuT phage typed as other than DT104. 

TABLE 4.

Reliability of markers for detection of multidrug-resistant S. typhimurium or S. typhimurium DT104

Isolate groupa	% Positive for:
	Chloramphenicol resistance	Florfenicol resistance	floSt-positive genotype
All S. typhimurium isolates with R types ACSSuTF or ACSuTF (n = 62)	100	94	97
Only multidrug-resistant S. typhimurium DT104 (n = 44)	100	98	100
All other Salmonella isolates tested (n = 47)	23	6	4

^aF, florfenicol. 

Plasmid pLB-510 contained a 3.3-kb fragment of DNA from S. typhimurium DT104 isolate 152N17. This fragment conferred chloramphenicol and florfenicol resistance (data not shown). Plasmid pLB510 was negative by both PCR and colony blot for the integrase gene (int). The 3.3-kb fragment was digested with PstI and EcoRI to form 1.2- and 2.1-kb fragments. The 2.1-kb fragment was strongly associated with both chloramphenicol and florfenicol resistance; however, no correlation between the 1.2-kb fragment and chloramphenicol or florfenicol resistance was found. Of 64 chloramphenicol- and florfenicol-resistant isolates tested, all were positive by colony blot hybridization for the 2.1-kb fragment. However, of 42 isolates tested which were sensitive to chloramphenicol or florfenicol, 41 were negative by colony blot hybridization for the 2.1-kb fragment.

We sequenced the 2.1-kb fragment from the insert DNA of plasmid pLB510. We found within the fragment an open reading frame of 1,202 bp. We examined that DNA sequence for identity with other published gene sequences by using the BLAST search function located in the NCBI database. The sequence had the greatest identity with flo_pp, a 1,122-bp florfenicol resistance gene described in Pasteurella piscicida (14). There is 97% identity between the DNA sequences of the gene found in the 2.1-kb fragment and flo_pp. Deduced amino acid sequence homology between the two is 84%, but if insertions and deletions resulting in frame shifts are ignored, the identity between the two is 99%. Identity (57%) was also identified between the 1,202-bp gene and cmlA (1) which is a 1,549-bp gene that codes for nonenzymatic resistance to chloramphenicol. Sequence identity at the amino acid level was 67% between cmlA and the 1,202-bp gene from the 2.1-kb fragment. The gene cmlA has been observed as a gene cassette in numerous gram-negative organisms (1). Due to the high level of identity with flo_pp and the fact that this gene has been described in S. typhimurium, we suggest the nomenclature flo_St to describe the gene.

PCR and Southern hybridization of the 2.1-kb fragment support our previous conclusion that the 2.1-kb fragment contained a gene coding for resistance to chloramphenicol and florfenicol. The presence of flo_St in the 2.1-kb fragment was confirmed by Southern hybridizations with the flo_St PCR amplicand as a probe and by PCR analysis with the Flo 1 and Flo 2 primers (Table 1).

One hundred nine Salmonella isolates were tested. Sixty-two were flo_St positive, and 60 of these had the florfenicol-resistant phenotype while one had intermediate resistance (Table 3). All 62 flo_St-positive isolates were chloramphenicol resistant. All 62 flo_St-positive isolates were multidrug-resistant, and both DT104 (n = 44) and non-DT104 (n = 18) strains were identified within this group. Of 62 multidrug-resistant S. typhimurium isolates, 94% (58 of 62) had florfenicol resistance and 97% (60 of 62) tested positive for the flo_St gene (Table 4). All 44 confirmed DT104 isolates tested positive for the flo_St gene (Tables 3 and 4), and 43 were florfenicol resistant while 1 had intermediate resistance (Table 2). Only 6% (3 of 47) of other Salmonella isolates tested were florfenicol resistant, and only 4% (2 of 47) were flo_St positive, although 23% (11 of 47) of the other isolates tested were chloramphenicol resistant (Table 4).

TABLE 3.

Summary analysis of colony blot hybridization results

Probe	No. (%) positive
	Multidrug-resistant DT104a	Antibiotic-sensitive DT104b	Multidrug-resistant S. typhimurium not DT104c	ChlrSalmonella not typhimurium	ChlsS. typhimurium
floSt	44 (100)	0 (0)	16 (89)	2 (12)	0 (0)
int	44 (100)	0 (0)	18 (100)	14 (82)	18 (72)
invA	43 (98)	4 (80)	18 (100)	17 (100)	25 (100)
spvC	44 (100)	4 (80)	11 (61)	3 (18)	22 (88)
Total	44	5	18	17	25

^aS. typhimurium DT104 isolates with R type ACSSuT, except seven isolates with R type ACSuT. 

^bS. typhimurium DT104 isolates with no known antibiotic resistance. 

^cS. typhimurium isolates with R type ACSSuT or ACSuT phage typed as other than DT104. 

The gene invA was found in 98% (107 of 109) of the total isolates tested (Table 3). However, spvC showed some specificity for S. typhimurium. Seventeen percent (3 of 18) of non-typhimurium Salmonella isolates were spvC positive, while 88% (81 of 92) of S. typhimurium isolates were spvC positive (Table 3). All multidrug-resistant S. typhimurium DT104 isolates tested positive for spvC. However, 88% (22 of 25) of the chloramphenicol-sensitive S. typhimurium isolates were also positive for spvC.

DISCUSSION

Florfenicol is a fluorinated analog of chloramphenicol approved by the FDA in 1996 for the treatment of bovine respiratory pathogens. Previous studies have shown that bacterial isolates which were resistant to chloramphenicol were sensitive to inhibition by fluorinated analogs (2, 3). Neither chloramphenicol acetyltransferase (cat) genes nor nonenzymatic chloramphenicol resistance (cmlA) genes provide resistance to florfenicol (1, 3, 11). Salmonella isolates with cat or cml genes mediating chloramphenicol-resistance also occur, and these isolates are not resistant to florfenicol. Conversely, flo_St-mediated resistance to chloramphenicol also confers resistance to florfenicol. Our research shows that flo_St has become a common genotype of chloramphenicol- and florfenicol-resistant Salmonella. Because the same gene, flo_St, confers resistance to both of these antibiotics on DT104, use of florfenicol may compromise use of chloramphenicol in treatment.

While numerous studies have reported on chloramphenicol resistance, only one previous study has described a florfenicol-resistant bacterium. This isolate was a multidrug-resistant isolate of the fish pathogen P. piscicida (9). Possibly due to the lack of previously published data in this area, the identification of 61 Salmonella isolates with a florfenicol-resistant phenotype suggests that florfenicol resistance is emerging in Salmonella. Interestingly, florfenicol-resistant E. coli isolates we have tested are also flo_St positive (unpublished information). The previous report (9) and this observation suggest that florfenicol resistance is likely to be described in other bacterial species. It is interesting to note that all DT104 isolates tested have been flo_St positive; this group includes isolates from cattle, swine, chickens, turkeys, dogs, horses, cats, etc., indicating that the use of florfenicol in cattle since 1996 may not be an important selective agent for this genotype.

Seven of the 44 multidrug-resistant DT104 isolates evaluated (Tables 2 and 3) were streptomycin sensitive. These S. typhimurium isolates were phage typed as DT104, indicating that the R type ACSuT may also be characteristic of DT104. The streptomycin-resistant isolates were also florfenicol resistant, and they tested positive for the virulence determinants invA and spvC. No evidence suggests that streptomycin-sensitive DT104 is less virulent than other S. typhimurium DT104 isolates with R type ACSSuT.

The invasion gene operon, invA, is essential in Salmonella for full virulence; it is thought to trigger internalization required for invasion of deeper tissues (4). Salmonella virulence plasmid, spvC, is associated with an increased growth rate in host cells and interaction with the host immune system (6). Ninety-eight percent of the isolates tested were positive for invA. A similar result was reported in a study where 245 Salmonella isolates from poultry, wastewater, and human sources were all positive for invA (19). The same study found only 37 of the 245 (15%) Salmonella isolates positive for spvC; however none of the isolates were serotyped, which makes a comparison of these studies difficult. Interestingly, the high percentage (88%) of S. typhimurium isolates that were spvC positive in contrast to only 18% of non-typhimurium Salmonella isolates suggests that some relationship between spvC and S. typhimurium may be present. Further investigation is warranted to determine the extent of this relationship and the effect it may have on virulence.

Integrons capture and express mobile genes known as cassettes, which are, in most cases, antibiotic resistance genes (13, 14). The integrase gene is an essential part of all integrons; it encodes a site-specific recombinase that catalyzes the insertion of gene cassettes into the integron. The results shown in Table 3 indicate that integrons are common in Salmonella, as 94 of 109 (86%) isolates tested were positive for the integrase gene (int). However, sequence data from the 3′ flanking region of flo_St indicate that it is not contained within an integron; this agrees with the findings of Ridley and Threlfall (15). Integrons have been documented in other gram-negative organisms, such as E. coli, Pseudomonas, and Shigella (10, 12). Therefore, it is unlikely that integrons will serve as useful tools for identification of Salmonella DT104.

Phenotypic identification of Salmonella isolates based first on serotyping as S. typhimurium and secondly on phenotyping as florfenicol resistant gives a simple method of classification as presumptive positive for DT104. Genotypic identification by PCR may be even quicker, as an isolate containing flo_St and spvC would also be considered presumptive positive for DT104.

This study has described a gene, now called flo_St which confers resistance to both florfenicol and chloramphenicol. Additionally, rapid phenotypic and genotypic tests for presumptive identification of multidrug-resistant S. typhimurium DT104 have been described.